Calculating the ITA Index
This morning I was searching through ITA Wealth Management to find information on the customized benchmark, ITA Index, and realized how little is available for the interested investor. Lack of information spurred me to write a bit more about it, why it is useful, and some of the potential flaws related to its accuracy.
Motivation to come up with a customized benchmark stems from the problem of comparing all types of asset classes with one single benchmark such as the VTSMX index fund from Vanguard. The VTSMX serves as an excellent benchmark for U.S. Equities, but what happens when bonds, REITs, emerging markets, and commodities are part of the portfolio. When this is the situation, and it is for most of us, the VTSMX does not give us a complete picture of how well our portfolio is performing. It makes no sense to compare bonds with the VTSMX. It is an inappropriate benchmark.
To pick up a video-audio of how the ITA Index is calculated, go to this link or transfer to the TLH Spreadsheet Help page where it is item #9. What we are doing with the ITA Index is measuring how well our investments are performing by multiplying the performance percentage times the percentage allocated in the portfolio. When this was written, 15% was allocated to emerging markets so we multiply 0.15 x the IRR percentage performance for VWO. And we do this for every asset class held in the portfolio. The video/audio walks the reader through this process.
Earlier I mentioned there are some flaws in the ITA Index calculation. At least these are the ones of which I am aware.
1. When cash is deposited or withdrawn from the portfolio, the same amount of cash is added or withdrawn from VFINX and VTSMX so we have a very accurate IRR measurement for these benchmarks. The same is not true for the other asset classes such as REITs and emerging markets. To do so would complicate the spreadsheet exponentially. It is already complicated. Since little cash flows in and out of these portfolios, this is not much of a problem for the 60% of the portfolio held in asset classes other than U.S. Equities. Nevertheless, the problem exists and we need to be aware it is there.
2. If one of the ETFs we are using to represent an asset class was not invested on the day the portfolio was launched, we introduce performance errors. In the Bohr portfolio we use BND, VEU, VWO, VNQ, GSG, and RWX to represent the six non-U.S. Equities asset classes. ITA Index, to reach its maximum accuracy level, should have had all these asset classes represented from the day of launch. This was not the case so an error is introduced in the ITA Index calculation. Over time this error will damp out.
Despite these two flaws in the measurement of the ITA Index, and there may be more, I still think the customized benchmark embodies sufficient advantages to continue to use it. As the portfolios age, the IRR for the portfolio will approach the IRR for the ITA Index.